Hacker News new | comments | show | ask | jobs | submit login
Ask HN: What do you wish someone would build?
186 points by prmph on Sept 23, 2016 | hide | past | web | favorite | 526 comments
It's time for another go at this question; we had interesting ideas the last time. What do you wish someone would build, either for your personal use or for your business?

Edit: fixed typo

A paid Linux.

To get stunning UX design, upstream bugfixing and excellent marketing.

Let me explain myself. I love the levels of ergonomy and polish of Mac OS X. But it's closed-source software. If I use (and pay) Ubuntu, then great patches are sent upstream, which I can use in Debian on my servers and Arduinos. It becomes useful to everyone. With Mac OS X, we're not advancing the world. But when I used Ubuntu for work, I was impaired compared to my colleagues. Blame it on a lack of seniority, but a steep learning curve for my OS is the last thing I want at work. So no Ubuntu, no elementary.io, nothing that has rough edges.

What allows Apple to hire UX designers and do bugfixing is the revenue. Which in turns gets them a good marketing team, which persuades the world of adopting their software. Linux misses advocates towards B2B, B2C and B2Gov. If we want adoption, we need a stable income, to improve UX patterns, bugfixing and marketing.

The FSF says it's ok to sell open-source software, but you just can't prevent people from redistributing. So it's possible to design a system that requires a yearly fee to access the upgrade repositories. Of course hackers will find ways around, publish a torrent, or choose not to upgrade. But the majority of people want a system that "just works" and have money to put down for this service. Businesses, programmer shops, owners of Teslas and iMacs don't want to download their OS from an unsecure source: They want the top-of-the-art, official, upgraded releases.

My own threshold is €200 per year for my work computer. We pay that much of IntelliJ. The OS is the most important service in our stack, it deserves paid workers. My parents' threshold would probably be 50€/yr.

I think OSS volunteers will feel cheated at first sight, but the software should really push changes upstream and show the value in having a much bigger Linux community.

NB: For those curious, this comment has 18 points so far (10:43 GMT).

You don't want a paid Linux. You want a paid Linux desktop-system - aka a "desktop environment software package" which support running on top of 3 most popular Linux server distros: RHEL, Ubuntu Server, Debian.

The only reason I and a lot of other people use Linux on desktops is because everything we work on runs on Linux servers and we want the same OS on our machines that runs on the server (even if we also use virtualization & docker when needed).

Even if someone made a perfectly polished Linux desktop system I'd still use some Ubuntu Server with an XFCE DE thrown on top because I like developing on a system most similar to what my software runs on. Incidentally, if I need to develop Windows software, I develop it on Windows. Only exceptions to this are Android and iOS since I couldn't find any good enough dev env setup for these systems.

If you ask for a "polished Linux desktop not based on a server-Linux distro", you are putting yourself into a very narrow niche.... a niche too small be worth developing for :)

> The only reason I and a lot of other people use Linux on desktops is because everything we work on runs on Linux servers

That may be true, but there are also a lot of people (myself included) that just prefer a Linux OS. I like having a choice of desktop environments. I like that a huge amount of software is freely available and just a "suda apt-get install" away. I like that there's no shitty bloat-ware on my systems.

I truly think this is a great idea. Literally the only advantage imo of OSx over Ubuntu to me is beautiful UX. Theres so many little tweaks and customizations. Everything else in Ubuntu is superior imo so I still use it.

Or maybe even an 'easy to use' theming system?

I know you can change window managers but thats technical and dense. It would be nice to have a way for designers and front end devs to mess with the UX.

I know the Ubuntu and Linux team have way too much going on but this a wish post lol.

200 bucks for a well UXed OS with support I would pay that tomm!!

> I know you can change window managers but thats technical and dense. It would be nice to have a way for designers and front end devs to mess with the UX.

Not really, you're a few clicks a way from installing KDE, Gnome or XFCE on any Linux and you can start hacking on it after some looks at the docs. Nowadays there should be plenty of HTML & CSS inside both Gnome and KDE themes... you just need the free time to waste on this!

Problem is not making the most gorgeous looking and butter-smooth animated desktop environment. Problem is you can't freaking expect drag'n'drop two work between any 2 applications because they are all so different and incompatible, or you can't have nice experiences with anything that need to integrate with the file manager because you don't just have one "windows explorer" or "finder", you have Nautilus, Dolphin, Thunar etc. And the zillion things that "half work": like, try copy pasting a folder open by ssh in Nautilus file manager... the damn idiotic thing will do a round trip through your machine and back instead of sensibly translating your operation into ssh commands...

A Linux Desktop Environment I'd pay for wouldn't have more diversity and hackability, it would have less, but all possible UI interactions and inter-program integration would be thoroughly tested and debugged. Also things like hot-plugging in/out your machine into displays and projector. All UI things that now work would work reliably! And customizatons would be a simple right-click/cmd+click + "options..." away on any UI element, not having to dig through pages and pages of settings in a damn "Control-panel-like" thinggy, or to install a buggy "Tweak tool".

As stable as Linux is for server applications, all Linux Desktop Environments I've ever used are unstable as fuck and I can get them to crash/freeze/delay-for-minutes probably once a day.

Ubuntu's unity had a right goal, but their head is up their asses - it's buggy, annoying to both power-users and new users, and pretty unhackable/obscure for non-professionals underneath. I don't event have a simple on/off toggle for "group windows in launcher task bar" which is 2 clicks away on Windows.

Their vision but "done right" and with enough respect for "regular power-users" ("power-users" who don't want to edit config files or install buggy tweak tools, that just want the advanced options baked in and thorughly tested!) is something to pay for...

I see what you're saying and agree. Just a system that makes everything compatible on the desktop so I can focus on my work without having to go through a major hassle to accomplish things.

Just a minimal bug desktop user experience with a reason amount of configurability for power users and if you want to go deeper there's always the command line.

I have master in graphicdesign/ux, i am ok with tech so i started to do programming a lot (to the point that economicaly its much better for me to just program than design). I am also big on open-source.

There is huge problem with open-source design. The kind of design you are talking about needs to be centralized. Design is about bigger picture, about consistency and overall aim. There are so many design solutions to achieve the same thing... usualy what ends up happening is that the dev implementing it just does it in his way. When this piles up it harms the project. I understand the dev implementing it, he has all the right. But overall its realy hard. And its a social problem, not technical one.

Would something like Red Hat Enterprise Linux (RHEL) or SuSE fit what you're looking for (both are paid, commercial Linux distributions), or am I misunderstanding?

I've thought about this phenomenon a lot:

The problem is that if you start giving your money to some group that could be working on the things that are most important to you, but they aren't working in that direction yet, then your support is understood to be a mandate for what they're already doing. Waiting until they're actually doing the things you want them to do before you start sending money their way doesn't work, because the point is to be the thing that incentivizes them to move in the direction you want.

This isn't so subtle that it's impossible to talk about and be understood, and it might work if your case is one that involves you being a patron for an individual or small group (e.g., through Patron). But for large, shambling, Red Hat-sized organizations, the message is one that's nuanced enough that you can pretty much count on it going uncaught.

But even for small groups, there needs to be a way to send the signal "please don't take this to mean that you can't stop doing the exact thing that you're doing now or else you'll disappoint; by all means, experiment on new stuff".

I agree with the patent that a highly polished version of Linux would go a long way toward further adoption.

As someone who has worked as both a sysadmin and software developer for many years, RHEL and SuSE are not any better than the other Linux distros in terms of polish and usability.

RHEL/SuSE are "enterprise" Linux, meaning you pay the vendor for a phone number to call when things don't work according to the documentation.

As primarily server-based operating systems, they receive no extra UI polishing than their free counterparts (e.g. Fedora, Ubuntu). I would argue that they often receive even less because as server operating systems, the vendors expect you to run them mostly in a headless environment.

In fact, since they are so "enterprise" and have long support/release cycles, they frequently featured software which is years out of date even on the day of release. Because the software has had several months/years since release, all the obvious show stopping bugs have been quashed, so now it's "ready for enterprise."

I would say that back in the day (mid-2000s), I would probably have considered Mandrake Linux to be the most polished. They had a graphical software center years before Ubuntu did. Unfortunately they found, as many others did over the years, that unless you're #1 or #2 in the desktop Linux distribution list, eventually the money dries up and you're toast.

Not to knock "enterprise Linux" too much. It has it's time and place: when you need a stable operating system with vendor support that you plan to develop and release a product on and expect that product to work for several years without major attention.

But for polished desktop use, they're worse than the other free distributions like Ubuntu/Debian/Arch/Fedora, both because enterprise Linux distributions lack recent software, and because the documentation is frequently behind a pay wall and only produced by the vendor (instead of say the Arch Linux or Gentoo wiki).

I remember RedHat as ugly. It would be closer to elementary.io, but paid. Well, the installation process of elementary.io displays unfriendly information too... At the extreme, choosing Mac OS X requires very few decisions, so I don't want to compare Linux distributions either. One distrib should stand out for my usecase, be clearly differentiated from the rest of the market and be the obvious choice for me. (that's where good marketing proves useful).

I know, it takes 1000 customers at 200€/yr to hire the first developer... Too much, too little.

SO, is your issue with a decent windows manager? I work on RHEL systems daily, and under the hood (sysd arguments not withstanding) its no more or less clunky than any other 'distro' out there. I guess I'm just trying to figure out what exactly you're looking to buy?

Yes it could be summarized as a window manager if it takes over all aspects of managing the computer through a GUI.

I develop with Mac OS and deploy my cloud services on Debian. I don't know much about Linux beyond apt-get upgrade, nginx, Ansible and Java apps. I know when I install Mac OS X that the progress bar is gray with no text on the screen, which means I don't need to dive into the technical details at any point from installation to writing a presentation, displaying it on an external screen through HDMI, tuning a mouse or plugging a printer. Has RHEL improved that much?

On step further is of course being able to install a good clone of Keynote (slideware), iMovie and an image editor for $40-80 each, but that's after the ecosystem starts gathering around the OS.

There is this great UI stack that is open source, has the latest and greatest animations. Its API is even used by thousands of developers already. It's called Android. Maybe it's time to ditch X11 and move to a more modern stack. Problem solved.

Ps: during the transition, you could support both. And of course the touch aspect needs to be cleaned up but it would still be a much better codebase to start from.

Have you seen Elementary OS? It will suit you if you are a mac person who doesn't really like the no-frills of other linux distros.

Yes I have. The installation process shows a progress bar in an old-school bevel/embossed window, underlined with the filename it's writing. On my computer it showed alerts in the middle. Once started, it's just a normal Linux with sharp edges. Installation of at least one of my programs (probably IntelliJ) crashed, I think I succeeded to fix that using advice from Stack Overflow and the command-line, if I remember. Definitely working, but definitely not the experience we'd like.

It's already a great OS, to be honest, but what about adding 20-100 employees and making it a blockbuster?

When was the last time you checked it out? That's what the team behind it has been doing. From what I understand the last two where all just polish.

I'd love it if you gave it a go and told us how you felt the pain points where.

I'm not involved with the project I just want more users on Linux-based OS.

I think the OP is thinking more from the general consumer perspective. Not sure whether RHEL or SuSE targets this market.

Interestingly, I had similar thoughts years back - https://www.quora.com/Has-any-company-tried-the-Apple-busine...

I want this, but added to feature list I want it to have certified-bulletproof compatible laptops. Pick some really good, top of the line laptops like the excellent current Dell XPS series or the MacBook, and make the OS work flawlessly with them. Drivers, touchpad settings, battery management.

The have that certificstion system thats supposed to give you that info?

Apple is not really paid for the OS, but for the hardware. Selling an OS seems nearly impossible today.

"We can't do it because it's a different business model than Apple's" ;) ?

From 9Gag to HN, Windows 10 is a recurring meme of disrespect for the -cattle- consumer, in a world of Snowden and Facebook where we don't even own our computers. I say there's a demand (with money available) for a trustworthy OS in the PC world, and for a open-source in the Apple world.

Selling today's OSes might be impossible indeed. But a lot of people would switch to Linux even if it had fewer features, just because of Linux' values (which would be properly marketed around values of compatibility, open-source, reusability, privacy, cryptography, distributed services, ownership, offline work, etc). What about marketing the idea that Linux will bring pervasive effects in our democracy, like a global understanding that security systems can only be verified if they're open-source? What about the community value of VW and Tesla being required to open-source their security features, generating better security for them, for competitors and pull requests to upstream projects? We can make customers dream about a lot of values in addition to adhering to a great-UX Linux, and this value is not currently captured by neither Microsoft nor Apple. If only we had the money to kickstart a paid Linux...

Have you actually tried to talk about those supposed advantages to non-tech people? Try it. Even many tech people don't care about that stuff and get annoyed if you start bringing up this kind of democracy, privacy etc. stuff.

People care about features and convenience. Getting stuff done and being entertained.

Seriously, talk to "normal people" and see for yourself.

When I created my current startup, all product managers and bankers told me it wouldn't work. That was 3 years ago and I'm still living off it. (Can't disclose the details of the product because I've disclosed things that could be linked to my partners previously with the same account)

So I'd bet people have been bitten enough by Android, Facebook, LinkedIn and Windows 10 that they're ripe to understand that their OS is worth 50$/yr.

What do you mean by "bitten enough"? Concretely. That there were newspaper articles about problems with their data protection, or about some data leaks? People don't care about this. It's irrelevant unless it impacts their workflow right there and then. Not in the abstract, not about ideals like democracy and ethics. They want to get work done.

"Bitten enough", for Windows 10, is:

- the forced upgrade to everyone who didn't agree with the upgrade,

- the telemetry

- the OEM "drivers" that display ads/change 404 pages/install a CA certificate/slow down an otherwise good the computer/install the OEM's wallpaper/break the trust between the customer and their computer.

For Android it's having coarse-grained permissions, the fact that "Ok Google" means the mic is on all the time, having Dropbox suggest to upload every time you take a picture. iOS is quite good but it's a closed garden, non-USB plugs, the impossibility of mounting the iPhone and just put mp3s on it, and requiring iTunes for music sync, which tries to push you to use their iTunes Store. It's all related to phones, but the point is, people do realize that they're providers don't care about them.

I said "normal people", not techies. Normal, average users don't even know these things exist. They can't agree or disagree with an update because they have no idea what a given update does and have no expectation to compare it to.

"people do realize that they're providers don't care about them"

Again, I can just suggest talking to average users. They don't even think about this sort of thing. They are annoyed if they can't get things done. If the app freezes, or drops the wifi connection etc. They don't care about permissions. They want their stuff to "just work".

My sister doesn't know much, but asks me all the time "Do you think my employer can bug my personal phone?" My parents asked me whether it's safe to enter their credit card numbers on their computer. My parents asked me how to get rid of notifications of their OEM antivirus. People who don't understand technology, but they do understand security, privacy and malware.

You're reminding me that my parents (63 years old) and my sister might belong to a bubble made of 10% of the people. I'm highlighting that those 10% still make above 100 million people. There's definitely concern for privacy all across the board, but people won't switch to Linux because open-source software is ugly and dense today.

I'm not saying we'll sell a paid Linux to everyone. But good UX + respect for the user is still a huge market. The market will grow as the image of Linux improves and people will end up switching "because it just works better".

Talk to HN folks and see for yourself...

Nobody (as in the general population) cares about all of that, unless they have been personally bitten by it.

Someone needs to take some Linux and focus solely on the desktop environment. Make it more beautiful than OSX and Win10 and that may drive attention. That's an awful lot of work (not only the apps but getting the ecosystem to cooperate).

We've driven ourselves into the ground with our desire for free things. And we paid with our privacy, willingly or not. I think that reverting this trend will take a very long time now.

First let's not say "beautiful", it's associated too much with transparent glass panes and shiny transitions. Let's focus on "functional": Fewer technical details, more do-one-thing-well and designed experiences.

Second, people have paid enough with their privacy ("Ok Google, were you listening to me?") that they've started to understand that it's worth paying $50/yr for an OS.

Anyway I'm not a Linux expert, so I can't even tingle with a business plan. Someone needs to do that for me.

Isn't Ubuntu basically functional and beautiful? I mean i installed Ubuntu for my mother, father and my sister's laptop. ( All of who only use their laptop for either watch Netflix or word=processing etc ). They were quite happy with it. They have been using it for a year now.

I think it's both. People like the hardware, but most of them get it primarily because it runs OS X. But, Apple makes most of it's money off hardware margins.

Apple bundles their OS very tightly to their hardware, so it's not as if there's really a choice between getting the OS or getting the hardware. Even the OS EULA forbids you from running OS X on any non Apple brand hardware.

Also, Apple is older than many PC companies and have been using their own OS since before Windows existed.

To expect a new company to come out with something that is competitive with Mac or Windows is to ignore the tremendous ecosystems that took decades to build around them.

I think there is definitely a demand for a beautiful Linux desktop system, but it would be a tough ask.

A few open source projects have had success with the crowdfunding model, when something non-trivial is in demand. If there is a possibility to get someone with a history of stunning design and UX on board, I could see a ton of money being donated for that persons time to build a rival desktop system.

I have differend position. As designer is extremely hard to help on open souce. And i know many designers who think the same. Usualy it ends up with nice logo. Ux - every dev wants to do on its own.

That was Mandrake/Mandriva, yet the margins were slim and they went bankrupt.

There were a few of them in mid-2000's. Lindows/Linspire is the first that comes to mind [1]. SuSE also had paid versions [2] (the free download was delayed for some time after major releases).

[1] http://web.archive.org/web/20030207074123/http://www.lindows...

[2] https://web.archive.org/web/20040803022103/http://www.suse.d...

Check out Elementary OS (https://elementary.io/). The desktop environment is very polished, and the hardware support is also excellent since it's an ubuntu derivative.

fwiw, I pay for RHEL7 Workstation ($299/yr for standard support, I think it's cheaper if you buy self-support) and it's fairly well polished out of the box. Also, now I have someone to call if I have issues. Not that I couldn't fix them myself, but I'd rather spend my time on issues I get paid to fix.

I would pay for this.

A thousand times this.

Isn't that pretty much the Mint business model?

A single IM platform through which everyone can talk to everyone regardless of their IM service, and I mean I wanna be able to send a message to someone's iMessage from my Battle.net account, and then receive someone's Facebook message on my Slack or something. Obviously IDK how this would be possible, but IM is now broken beyond repair by companies that tried to "fix" it. The list of IM apps on my iPhone keeps growing, I got Slack, Hangouts, Telegram, FB Messenger, iOS Messages, HipChat, and Skype, and whenever I need to search for a message I never know which one to look in, and whenever I need to message someone I never know which channel is the best to use, and it's just a freaking mess. I hate state of IM in 2016, and I hate the parties that were involved in getting it to where it is now.

I guess some of the semantics I've mentioned in my comment were wrong. I don't care about Battle.net, Facebook and Slack. All I want is to open ONE IM app, enter friend's name, send them a message, and receive message from them in the same place. And I don't want to know who's on Facebook and who's on Yahoo Messenger. I don't even wanna see those icons next to people's names or anything. And I want it to have a web client. And amazing search. There goes your billion dollar company.

If there was a billion dollar company to build on this premise, someone would have. In practice, users don't care about having to install five different messengers and not using their own clients.

Do you suggest that all billion dollar companies have been already founded and there's nothing new which can be done? I find such way of thinking harmful.

No, I suggest that all billion dollar messaging companies that don't lock their users in have already been founded. I would love nothing more than to be proven wrong.

This. It's baffling; we've "solved" email – anyone can have an account with any provider and is able to mail anyone else – why haven't been able to do the same with instant messaging already?

Because email was already standard when internet became commercial; if you wanted to be part of it, you had to have some email infrastructure set up.

Today no company needs any IM, it's only a product (instead of a communication protocol) geared towards the masses. There's little business reason for them to use a standard protocol.

Then let's just use email as the underlying infrastructure to message people! You would just need to build a ui to make it look like you were chatting with someone :)

I was thinking about the same thing and just posted a comment to the same effect :o


You would think it would be a no-brainer that everyone expects to work as it basically works to reach "everyone" in the world with far older technologies like physical mail, E-Mail and the telephone. We even have a standardized protocol (XMPP) to interoperate between IM services but it's largely unused because users mostly don't demand such a feature and big providers are better of locking everyone into a walled garden.

An open protocol can also lead to an inconsistent user experience. In the year 2016 I still have no idea how to send an image from one XMPP client to another and be sure that the sending side will receive it. In modern closed-off systems this is braindead simple. Part of this is due to how fragmented XMPP is, which I would again argue is due to its openness.

> An open protocol can also lead to an inconsistent user experience.

No more so than a proprietary protocol. All the IM clients and apps are different, just like all the email clients.

What matters is that I should be able to just get anybody's IM address and message them with my client and service of choice.

> No more so than a proprietary protocol. All the IM clients and apps are different, just like all the email clients.

What? No, if my colleagues and I all decide to use ${CLOSED_CHAT_SERVICE}, the user experience is generally going to be consistent.

> What matters is that I should be able to just get anybody's IM address and message them with my client and service of choice.

This is what matters to you. I don't care what client I use if I can't use it in a way that meets my business needs, like sharing screen shots.

> What? No, if my colleagues and I all decide to use ${CLOSED_CHAT_SERVICE}, the user experience is generally going to be consistent.

But even in email, if a group of people decide to use only Gmail web interface for example, the user experience is going to be consistent for them as well.

The underlying detail about email being an open protocol will be transparent to them.

In fact, an email client could be designed to implement instant messaging — by sending messages as emails — and if two people both use that same client, the experience would be indistinguishable from an IM service!

You'd still have a conversation history, you'd still have offline delivery, you'd still be able to send graphics and animations and audio attachments and anything else that HTML can render.

(I get that current email protocols may not be lightweight/efficient enough for rapid delivery of a large amount of short messages though.)

Because of the spam threat. Controlling identity (for example by knowing for sure the sender's phone number) allows for a huge mitigation of spam.

What's different about email?

In email you have sender impersonation. Once blocked, adding a new sender (cell number) is expensive. It is 12 orders of magnitude easier to spam with email than it is to spam with whatsapp.

But that's exactly Razengan's and my point. The email spam problem was solved even though email has very little identity verification. So spam can't obviously explain why instant messaging systems are so walled-garden incompatible, but email functions fine.

Email protocol was by fiat. We don't have that luxury anymore We have the free market. Hooray!


matrix.org is trying to do this, but so far progress seems slow/incomplete so far. No comment on the technical aspects of it, but the idea of my own server where I can run bridges to all these different accounts is pretty appealing. If all these bridges were actually available, I would be on it all the time.

Ive been using it sucessfuly for over a year. Its moving forward quite well i think.

On my windows phone 7 I was able to seamlessly message my friend via text or Skype or Facebook message, all in the same conversation thread, which was searchable. and when going on his page see his Facebook and twitter updates in one page.

From Wikipedia:

> Windows Phone 7's messaging system is organized into "threads". This allows a conversation with a person to be held through multiple platforms (such as Windows Live Messenger, Facebook messaging, or SMS within a single thread, dynamically switching between services depending on availability


too bad windows apps required a comprehensive bottom to top knowledge of windows stuff. for example - built an android app while learning java in school. didn't know what everything did but the outcome was the same as promised in the tutorial meanwhile in the windows 7 app tutorials, phrases like "do this as you would do in wcf so & so and it should work" were thrown about leading to an endless nesting doll structure of learning what the definitions meant- that led to 4 weeks of on again off again learning before i made a copy of my android app. the tutorials improved later but i'd given up on windows app development by then, the promised free support for windows app developers never materialised in my country

great platform decisions, poor app developer support and shitty ecosystem(still haven't considered developing for the windows 10 appstore because of this reason).

What happens when you send a message? Does it spam to all platforms? Does everyone have to declare a preferred app to receive? Does it check your previous communication with that contact to see which platform has the best response rate?

Just asking, it doesnt seem like it would be that difficult. In fact im sure this used to exist when it was just msn, icq etc.

I would imagine that, quite like Email, you have a universally recognizable IM handle that you can use with any service.

Actually, in China, everyone uses WeChat. It is the ONE IM app you described.

All it would take for this is more love from the existing apps from big corporations (Skype, WhatsApp, etc) towards properly supporting mobile, desktop and web clients.

They already have all the users, which is really what makes a messaging platform (I won't use a new app if my family/friends/coworkers aren't there even if it has the best clients everywhere).

Skype seems to be trying with the new web and Linux clients but it's too little too late. There is still time to revert it through. If they opened their API so anyone could embed a client on a website, it would be massive.

That's not solving the problem in any way though. How do I send a message to a friend's iMessage from Skype?

Honest question: what's wrong with email? Not in the sense of "you should just email rather than IM and stop complaining". In the sense of "Why doesn't someone build a dead-simple overlay that exploits the email protocol?" I'm thinking a simple tag at the beginning of the subject line that causes it to be auto-archieved in gmail. Is the latency too large?

Yes, the latency is too large.

When I want really instant messages I've to switch to WhatsApp because Hangouts on my phone takes a minute or two to receive messages (tried all configuration tricks in the book). Just an example where latency with messaging apps is already too high.

Yea, Hangouts is garbage...

The fact that the identity of the sender is not verifiable is a pretty major flaw. SMTP is also too often unencrypted. And it is not designed for real time.

A friend recently started working for a company that builds the app pairade. It supposedly does that. Maybe you should check it out.

Disclaimer: I have never used it.

if you sum jabber+icq+yahoo+msn+etc.. you get probably 5 users. :(

It also supports anything that pidgin supports, which gives you WhatsApp, Telegram etc.

IIRC ICQ is still popular in Russia?

I am quite surprised you do not have WhatsApp.

A slim, networked, pocked sized computer with a physical keyboard, running android or Linux.

These used to exist - albeit disguised as phones - but the marketing department decided we don't need no friggin keyboards, and remove everything but the touchscreen and call it a tablet. The result is a consume-only device, on which it is all but impossible to input large amounts of text.

Nokia N900, Motorola Droid 3/4, HTC Desire-Z. These were the last of their species.

Like the pyra/pandora? https://pyra-handheld.com/

the problem with every single pandora-spin off is this: they try to do a full product and sell it like a finished product (mostly to game pirates)

What they should do: work on the case, and do something like the ergo-dox cases[0], where the case walls are a pile of 0.5" planes, so you can have as much space inside for your additions as you want. And then on the top of the case include the LCD clamshell and the keyboard/joypad. Because, honestly, that is the ONLY good thing on the pandora-clones.

Then let the insides of the machine be whatever the user/kit-vendor wants. Raspi, android all-in-one, etc.

that way you don't get outdated inputs (compact flash in 2013!) or crappy cpus (EU$300 for a celeron in 2014). The keyboard and the screen will mostly stand the test of time better, for example, 720p for 90s game emulation is almost overkill.

[0] http://adereth.github.io/images/case.jpg

I just wish that thing wasn't so much money. It's definitely worth it, lots of engineering behind it. It's still just a large chunk of change.

i don't care about the money, i have and will in the future spend lots of money on useless things what i'm always worried about with these projects is the potential lack of community engagement. tons of these type of things come out and die within six months because the major maker/hacker communities just do not give a damn about them, so they don't get iterated on or mixed with other things. makers/hackers are fundamentally cheap ie they like working with cheap components even if the overall scope creep on a project ends up eating a significant chunk of their salaries.

raspberry foundation did something really clever by going for the cheap/educational route. nearly everyone i know has heard/hacked on a raspberry pi

exciting to hear about and if the project makes it in terms of excitement within the maker/hacker communities, will definitely buy one

Depends on your perspective. What brought me around was comparing it to an iPhone - similar price, yet it addresses pretty much every standard nerd complaint about the mobile ecosystem (keyboard, free software phone, standard Linux). I felt that not buying one would be simply hypocritical :)

That's how I feel. Then again, I'm a college student who can't afford one sadly.

I like the idea, but they talk about the future while the website is designed like it's still 1997.

https://en.wikipedia.org/wiki/Psion_Series_5 form factor with modern phone hardware.

Yes, this, exactly

A few days ago I came across a review of this:


Seems to be running on Android and has physical keyboard. Has anyone used this phone?

yes. it was so-so when launched. Then with every update it got WORSE. i kid you not.

for example, i will just list the worst case so you can have an idea: they updated the keyboard so that now the backspace also works as delete. But you have no control! it decides! You carefully position the cursor on the middle of the line, press backspace to erase let's say, the http:// part of the url because you want to type in ftp://, well, though luck, because as soon as the cursor reaches the beginning and there is nothing else to 'backspace' on the left, they key becomes Delete right under your finger, while being pressed, and now it deletes the rest of the URL... it is maddening.

But, it had so much potential. The keyboard has touch capacitance on EACH KEY! so you could in theory use swipe type on the physical keyboard. sadly, the swipe type from blackberry is just garbage. And if you select any other soft keyboard, then nothing from the physical keyboard works. No prediction row, no cursor control using capacitance touch. not even shift or symb keys in some cases.

Agreed. They don't make niche devices anymore.

The last great phones for me where the N950 and N900.

Everyone is building what they think appeals to the "average consumer". With every yearly product iteration, features regress more towards the mean. They just build bricks that look the same now, competing with each other on how many fractions of a millimeter can be shaved off. This is because phones are no longer seen as productivity devices, they're a status symbol like cars and jewelry. A keyboard increases thickness, which would be unthinkable for them.

I actually invested in a Jolla phone two years ago, but the promised hardware keyboard never quite panned out.

The market, not the marketing department.

>> Nokia N900, Motorola Droid 3/4, HTC Desire-Z. These were the last of their species.

Yes, but not the pinnacle of keyboard - Nokia 9300i for me...

  >> Nokia 9300i for me...
Yes! Yes! That thing was THE best phone I have ever had.

> Motorola Droid 3/4

I recently and reluctantly retired my Droid 4, simply because Android 4.1 was getting old (no stable newer versions available) and the phone was too slow for newer apps.

Now I'm on a Blackberry Q5 - Purely for the physical keyboard. The downside is that I have to live with very limited app support :(

Well it's kind of a geek's toy, not a productivity champion, but there's PocketChip.

I totally agree. I used a NEC MobilePro 780 & 790 for a few years in the early/mid-2000s and it was awesome. It was small enough to slide into my pants pocket, yet still had a quite usable keyboard.


I'd love to have something like this running Android or even a very stripped-down Chrome-like OS.

I loved the Palm Pre's form factor. The size, the sliding hardware keyboard, the way it felt like I was holding a smooth stone in my hand. It's probably the best phone I've ever owned, hardware-wise.

The manufacturing quality was abysmal, though. The first one I got had a defect whereby sliding the keyboard shut would cause the device to turn off. The second one I got after returning the first one had a defective screen.

Will this not be solved, partly and in a sense, by hologram keyboards? It seems your wish for a physical keyboard is not due to an enjoyment of the tactile nature of it but rather that it's quite hard to type on an in-screen virtual one.

Hologram keyboards would still seemingly require a surface to type on, though, so it wouldn't solve everything. One might imagine a future where the phone is in your pocket but a hologram screen + keyboard appears (somehow) in front of you. I'm just thinking out loud here.

You can't hold a phone by its holographic keyboard. That means you'd have to type with one hand. One of the big advantages of physical keyboards is the ability to type with 2 thumbs, or even 2 hands on devices like the Psion 5.

You can plug a usb keyboard into any Android device and it works fairly well for input. I can use an ssh program like ConnectBot as if it were a regular terminal with an attached keyboard.

And deny yourself the satisfying, visceral snap of an integrated physical keyboard sliding out?

Using a USB/bluetooth keyboard is certainly functional, but it's an extra item to carry around, and it's not as elegant.

To be more specific, I would like a BlackBerry Passport running Android or Linux. Additionally the keyboard should carry little LEDs so that switching keyboard layouts is easy.

I bet you can put Linux on the Sony Vaio P, though it's pretty dated hardware at this point

I hypothesize that the list of ideas that will be posted here will make for great examples of what not to build if you're considering commercial value. Developers and the HN crowd make for a very small market and are very hard to monetize.

In that vein, I wish someone would build a list of things that regular, everyday folks actually would want and use. The middle manager working at BigCorp; the teenager; the stay-at-home mom; the retired; people who actually want to spend money to solve their problems.

Thought of this idea as a platform; Did market research; Not enough people think of their own problems in solvable terms; Also a lot of problems are too small ("mosquito bite") or general complaints about the situation ("it's too hot in southern France").

Might work as a curated list, but then it is hard to scale.

This was linked a while ago: https://nugget.one/

Not affiliated, not even a paid customer, just thought it was well made.

1) More high quality news analysis content. Think NASA Earth Observatory, The Information, or the best articles you've ever read, and put them behind a paywall. Consistently making top notch content is hard, but I suspect it can be easier if writers are paid good wages to explore their interests and the news industry decouples itself from advertising. I'd pay for it, and I suspect that over time, enough people would.

2) A personalized learning resource. None of this AI adaptive learning nor passive MOOC lecture watching. Get people who know what I want to know (usually job-related skills) and have them sit down and teach me things 2-3x a month. I want structured, supportive, long-lasting mentorship from people who genuinely want to see me grow.

3) A doctor that proactively cares about my health. I hurt my shoulder, but aside from a 30 min physical therapy appointment once every other month, I'm on my own. My posture sucks despite having a split keyboard, standing desk, and doing exercises to fix weak muscles. I need someone to make me diligent about my own well being, day after day. I want workout buddies. I want someone who will pick up yoga just so I'd have someone to do it with. I can get more health benefits from a concerned friend than a licensed medical professional.

4) Life training. Working in tech makes me feel detached from humanity. I want to be a more loving person (I am one, but the culture distorts things and makes me think about my skills/career/startups/work/money too much and life too little). I want someone to help me take 8 weeks off a year to spend time with family and go on vacations. I want someone to help me be a better parent when I have kids (my parents aren't great role models). I want someone to remind me to appreciate all the things I have in my life.

What products I use, how I store my data, etc are just sweating the small stuff. Health, education, happiness, sense of community, etc -- fix my big, recurring problems that truly matter to me as a human. Go above and beyond to do so and pay way more attention to detail than most software products do today. Relentlessly follow up on everything. Keep it human and personal.

These are probably not the answers you wanted to hear, but these are needs that grow bigger and are usually unaddressed over time.

3, 4, and the emotional support aspect of 2 read like you are pining for matrimony.

haha perhaps matrimony would address some of the issues (I am starting to approach THAT age). I'm just a big picture guy and over the years, I don't think that tech has helped that much in addressing some of the more fundamental life problems. Some exceptions (skype to chat with relatives abroad, facebook/social apps to a small extent, my hospital finally creating a website, youtube fitness tutorials, etc). Tech can totally help solve these problems, but there's a ceiling I hit on all of those categories that require me to continuously exert a high amount of willpower to break.

What I want someone to "build" is the cultural environment to make it easier for me to be proactive about my wants. I want a sense of community with my city. I want friends who make themselves more available (and to be one myself). Maybe these are the things that an idyllic small city supposedly offers that big cities do not? I don't know why people don't really talk about these topics often, but maybe it's just me.

Large cities rather self-consciously (as any NYC comedian will tell you) run on the razor's edge of minimum comfort for the maximum amount of productive people, ruthlessly filtering for those who can withstand the continual mental, physical, and social penalties with vitality to spare.

I don't want to sound preachy, but having felt nearly identically to what you've described in #4, I can tell you that I've found my answer in meditation. It's allowed me to slow down life and realize what's really important. It's hard to get caught up in those ego-driven things now that my sense of self and ego has so inexorably changed.

It has also helped with my posture too.

glad meditation has helped ya -- i think everyone gets different things about it. It helps me slow down and be more present instead of worrying about all those career things. It makes me feel a little more connected to the world. It helps me reflect on who I want to be.

I am working on #2. There has to be some sort of incentive for the mentors for this to be sustainable. As a mentee what can you provide to the mentor? What do you think will make this relationship be sustainable?

This area is interesting to me - both to learn and teach. In the past I've explored proffered solutions and found them too heavy and too slow. I'm interested in helping make something that has low friction and is more peer to peer - ping me if I can help but I'm sadly reticent to take the lead.

I feel like the best mentor/mentee relationships aren't based on incentives, but a friendship/concern for one another. A small handful of senior ex-colleagues who I've worked with genuinely look out for me and my career interests. It takes time from their day and often unrewarded intros to their network. Aside from goodwill (and maybe a referral bonus if it leads to a job), there isn't much incentive for them.

If we treat it as transactional, then certain relationships will never foster. For example, if some high school kid got admission to a university I went to a few years ago and wanted my perspective on how to make the most of his education, I have no discernible incentive whatsoever to answer (assuming he can't pay me). Giving a quick response is either useless or hurtful without taking the time to understand his needs and wants. But that kid could probably benefit tremendously if someone took him under their wing.

I think some people want to feel they have accumulated a lot of skills and experience and simply want an outlet to give back in a manner that feels effective to them.

Maybe universities can tell their alumni network that if people mentor new students, they get access to premium recruitment opportunities on campus? Or cities can give tax rebates to people who take an underprivileged youth under their wing. Employers can promote more externship/shadowing opportunites and give invites to the employees. Or maybe senior positions at companies should have "train junior employees" as one of their top job objectives, and alter promotion criteria to bias towards that.

As I've been told by people who helped me, you don't pay back, you pay forward.

When you're young, you need the help. When you've grown thanks to that you give it forward to the younger ones.

For #4 check out The School of Life: https://www.theschooloflife.com

> doing exercises to fix weak muscles

What kind of exercises? If you're not doing e.g. deadlifts, rows, facepulls, squats, etc., you're likely wasting your time. There's no better way to increase overall stability and muscle tonus than hypertrophy. And there's no better way to drive hypertrophy than progressive overload with weighted movements.

I was physically most at risk after I had put on 50lbs of muscle and was routinely breaking lifting goals:

Too heavy benchpress / lifting under fatigue to promote growth --> high risk of accidents.

Too little stretching and lots of heavy squats /deadlifts made my back very tight and prone to injury.

A workout biased towards a high back:chest ratio didn't offset all the time I spent at a computer, so my shoulders would round forward but now my muscles/tendons are even bulkier and have more friction with my joints

I realized if I'm not strict with correct form, range of motion, rest between sets, proper progressions, stretching, etc, hypertropic workouts are more dangerous than beneficial. And if I am strict with everything, I can recover and get stronger at a much healthier pace even if I never lift more than half the weight I used to (just slow the tempo down). I might not get as muscular as before, but I got no benefits from being bulky.

I have a very similar experience with #3. I feel like so much about care for chronic injuries is super sub-optimal.

Almost reads like a list of reasons to move back to mom's house.

Sports (and events) Netflix/live streaming. That would kill cable for sure.

A curated "channel-like" experience for Netflix/YouTube shows. There's a LOT of good content out there, but it's hard to filter. And the "channel" experience of surfing and switching between programs has been kind of lost.

A different kind of smartphone, with actual buttons. Or maybe even what one manufacturer (Samsung?) tried to do, splitting the phone experience from the smart experience, with one ergonomically good device for calls, and another for messages and browsing. I also very often want to be looking at my screen while on a call (check mail, google stuff, look at maps).

I also miss the experience of the slide-to-answer on my old Nokia n86, or flip-to-answer like the Motorola Razr. I could also make calls without looking.

"Sports (and events) Netflix/live streaming"

The issue here is rights.

The rights to this content are very tightly controlled, it's very expensive, and contracts are allocated usually in the billions.

It would take a startup literally with billions to 'out bid' some entity for right to a sport, and then make it available.

FYI - if you subscribe to ESPN etc. for cable, I believe they actually do have streaming available, but since I'm in Canada, I can't say for sure.

It's not a technology issue, it's a value-chain issue.

A lot of entertainment markets are kind of screwed up because of this - and how arcane some of the systems are.

irt sports streaming: streamers/xmbc people have been doing that for a while http://koditips.com/kodi-pro-sport-addon-streams-from-reddit...

there's even news of a crackdown going on in the uk with people who are selling modded boxes http://www.mirror.co.uk/news/uk-news/kodi-illegal-first-man-...

Right. So I imagine the OP would like a reliable service and not have to keep playing whack-a-mole chasing unauthorised sources.

A lot of the 10ft apps available on e.g. Amazon Fire TV etc fall short here too - I don't want to keep selecting the next 3 min clip, just instantly play me a curated stream.

Mark Cuban had good thoughts on TV vs VOD many years ago e.g. http://blogmaverick.com/2010/10/26/the-value-of-your-time-an...

The issue is not technical, its the usual problem with audio/video content and vested interests - regional locks, contracts, existing cash cows etc etc. Tough for a startup to negotiate deals against Eddy Cue, I see the solution coming from the big tech Cos...

Last week a friend recommended http://www.dazn.com to me which aims to become "Netflix for Sports" (AFAIK available in Germany only).

On android you can select the home button to answer calls. And the power button to hangup. Assuming you have these it works nicely for tactile answer and end.

An operating system with an interface that's a substantial improvement over POSIX. I'd like to see statically typed files and pipes, and a built-in file conversion solver so that if I have, say, a postscript file and I want it in pdf format, I can just tell the OS to use any conversion utility with a matching type signature and I don't have to remember what it's called.

I'd like a process/thread to be able to have multiple current-working-directories at the same time, so that a library can change the directory without effecting the rest of the program.

Similarly, I'd like processes to be able to operate on behalf of multiple users at the same time.

I'd like stricter security policies that deny network access, file system access, etc.. unless they've been explicitly turned on (like Android or IOS).

I'd like transactional semantics for filesystem updates. No process should be able to see changes made by another process until that process does a "commit".

I'd like to have a general command-line undo. I'd like to be able to do "rm -rf /*" and then undo the operation and have everything be restored.

I'd like to have something like proc files, but for user space applications.

My full wish list is quite a bit longer, but that's good enough for now.

My item added to that would be this,

"There's a universal dedicated storage mechanism for the state of the operating system and its related packages similar to a SQL database."

It really annoys me how ad-hoc state and values are treated in an operating system with ASCII config files.

You can always have means of human-friendliness even if the underlying data is more strictly structured such as in a database but you can't go the other way around.

Imagine almost like a SQL-like language to provision/configure/restore an operating system.

   SELECT name FROM installed_packages WHERE installed_date > '2016-01-01 00:00:00';

   DROP FROM installed_packages WHERE package_name = "python2";
   INSTALL PACKAGE WHERE package_name = "python3";
Operating systems are still stuck with an over-zealous "the unix way or the highway" mentality of cobbling together fragile ad-hoc config files with arbitrary config languages.

Each package invents their own config language and forces it down your throat.

WinFS tried and failed to do something similar.

>Each package invents their own config language and forces it down your throat.

Windows Registry.

elihu, these are some good ideas here. I would love to see your "full wish list". Can you post it somewhere?

More improvements on the POSIX interface:

- Processes should be able to reserve some % of both CPU and I/O bandwidth. So you can have video players, games, etc. which remain perfectly smooth no matter what other processes are running at the same time.

- I would also like it if the kernel and its drivers would refuse to 1) write to the boot sector of any disk, and 2) reflash firmware on any device, even if requested by root, unless the computer is booted in a special "maintenance mode". That would eliminate persistent rootkits.

> - Processes should be able to reserve some % of both CPU and I/O bandwidth. So you can have video players, games, etc. which remain perfectly smooth no matter what other processes are running at the same time.

It is far from the nicest interface, but I believe this can be achieved with cgroups in Linux (see `man 7 cgroups` "Cgroups version 1 subsystems").

The cpu subsystem: "Cgroups can be guaranteed a minimum number of "CPU shares" when a system is busy. This does not limit a cgroup's CPU usage if the CPUs are not busy."

The blkio subsystem offers "a proportional-weight time-based division of disk [I/O] implemented with CFQ." Though this only guarantees a proportion of I/O, not a specific bandwidth.

Okay, I found the list where I wrote these down. Here's some more:

- The ability to make changes to /proc files that only persist for the lifetime of the process that made them. (Basically, turn proc files into a stack, for cases where some program requires a particular setting, but don't want to make the change permanent.)

- globally managed weak pointers. If multiple processes are managing large amounts of cached data that can be regenerated if it's thrown away, the OS should be in charge of decided when to keep the data and when to thow it away, as it has a better view of the complete system than individual processes. (There's been some work to add this to the Linux kernel, but I don't think it's available to user space. https://lwn.net/Articles/340080/)

- let's get rid of statically-sized partition tables. There's no reason they should be a fixed size on SSDs. (It may be necessary to set a minimum and maximum size, but within those bounds they should just grow and shrink as needed.)

- A regular user should be able to create sub-user accounts with a subset of their own permissions. This could be useful to run untrusted applications in a sandbox.

- It should be easy to suspend a process to disk, migrate it to another machine, and start it back up, or to kill it and have it restart in the same state it was in when it was killed.

- It should be possible to restrict a process from opening any files that weren't passed in as arguments on the command line. (I think this has already been done. I don't remember what the project was called.)

- Package files (like rpms or debs) shouldn't have fixed file paths in them, or make any assumptions that the local filesystem will be organized in a particular way. Similarly, package files shouldn't run scripts. Ideally, only the package manager would be able to write to /usr (or whatever the administrator wants to call it on their system).

- It may be possible to do away with the user-space / kernel transition, and just run everything in a single address space. The way to do this safely is to write all software in a language that doesn't allow shenanigans (such as Rust) and only permit the OS to execute binaries that have been created by a trusted compiler. If the compiler can verify that types and memory boundaries and API conventions are always respected, then having the hardware check it again at run-time is redundant and an unnecessary drag on performance. For cases where untrusted code blocks are necessary, those would behave like kernel code; you need to be root user to tell the OS it's safe to run such code. Legacy C and C++ applications can run under an x86 emulator or inside a partitioned-off address space, but performant applications would have to be written in a safer language.

For typed, convertible files, global revision control, and possibly the access control enhancements you mention (for the other features ignorance begets silence), Urbit [0] may be of interest.

[0] urbit.org

> I'd like to see statically typed files and pipes

You mean homogeneous files and pipes - only text, only numbers, etc - that kind of nonsense? Static typing falls apart with heterogeneous containers.

Or you are just putting that static typing everywhere because it is so cool - typescript, huh huh?

The idea of types is a little different for files than it would be in a programming language. In a sense, all files are trivially of type [Char], but that's not really what I mean.

If program A emits XML on stdout, and program B accepts JSON on stdin, "A | B" should be rejected by the shell.

A related feature that might be useful is to be able to register checkers or filters for any file type so that if, say, same application creates a pdf file but the data it writes out doesn't look like valid pdf data, it's rejected. Or, if a process tries to write out an mp3 file, the OS converts it to an ogg file to store on disk and converts it back if the application tries to open it later as an mp3.

Would this codify that .xyz filename "extensions" are mandatory.. or would a file's "type" not be stored in-band with its name?

For media files you would be able to achieve something similar with FUSE and FFMPEG.

A stack for building web applications in the browser. HTML and CSS are pretty good for documents, but terrible for in-browser GUI apps that we're all building, its just piles of hacks upon hacks.

I want someone, probably Google since they own both a major browser and some of the most popular web applications, to re-invent the entire stack. Steal ideas from GUI-focused languages and toolkits, like QML, Swift, AppKit, etc. Lets pull in a superior scripting language like Lua and widget-layout framework like Qt, and support it natively and securely in the browser, building on everything we've learned in the last 20 years of creating web applications.

An OS based on IPFS would give you the best of both worlds.

AFAIK browsers are popular because they load apps quickly and without an install step. While native apps let you choose your stack. An OS based on IPFS would load big apps quickly thanks to caching, the install step would go away because there is no difference between remote files and local ones and its still an OS so the app can be written in C or any other language built on top of it.

Zero-install is half of the reason why browsers are popular. The other half is that the Web puts the users in far more control over the content in their viewport than anything else that has been widely deployed except for probably Microsoft Excel.

If your first thought is that desire for control is a niche thing, an instance of power users projecting their biases onto the rest of the world and not a concern for some mythical "everyman", you are wrong. Because those normal folk are exactly why Excel even gets a mention.

That doesn't explain the popularity of mobile apps among the "everyman". They're not very customizable, but they're considerably more popular than either the regular or mobile web.

> If your first thought is that desire for control is a niche thing, an instance of power users projecting their biases onto the rest of the world and not a concern for some mythical "everyman", you are wrong.

Evidence? Sales numbers tend to prove the opposite. It's the whole reason the apple ecosystem makes as much money as it does: it just works and it looks good doing it. Most people don't enjoy having to spend any time on getting anything to work or customizing anything to their liking. They just want it to do what it's supposed to do then go back to whatever it is they find more important.

Now, could a zero-install network of apps kill the browser? I'm not so sure. Developing native apps is a royal pain in the ass right now. But, if something like QML were to get a lot better and easier to use, it could take a dent out of the single page application market.

> Most people don't enjoy having to spend any time on getting anything to work or customizing anything to their liking.

Well that's great, because those aren't words I ever said. The first person to start talking about customizing things is you.

I'm not going to be nudged into mounting a defense for an argument that I never tried to make.

I think even electron apps would be better than web apps with an IPFS OS to make it load quickly

Man, I'm late posting this, but this is exactly what we're building with Anvil - https://anvil.works

> Superior scripting language

We use Python - on the client and on the server (with rich RPC between the two).

> Steal ideas from GUI-focused languages and toolkits

The biggest such idea is that most layouts should be done visually - Visual Basic had this right, and we've been moving backwards ever since. (The web has three, usually four intermediate representations you have to think through to edit your appearance: Template (->substitution->) HTML (->parsing->) DOM (->CSS render->) visual layout)

That's what I understood Microsoft Gazelle[1] to be. I remember a blurb about it in 2009, I was very excited about the idea, but since then I haven't seen or heard anything.

[1] https://www.microsoft.com/en-us/research/publication/the-mul...


What if in the last 20 years we've learned that browser is not the proper place for applications?

I think we've learned the exact opposite of that in the past 20 years.

I disagree with that very strongly. I don't want to have to install 100+ desktop apps (and hope they're available for my OS, and up-to-date) to be able to use every service I interact with. If GMail, GCalendar, HN, GitHub, CircleCI, Google Docs, Slack, and every other web application I use every day were not in a browser, the world would be a much worse place. But all of those applications are built upon that pile of hacks, and much harder to write an maintain than they should be.

Why would you say that? The browser has become the most successful app platform ever built.

I would have said Windows was the most successful app platform ever built.

keyword is was. no one is excited about making innovative windows apps, it's a way to get your thing out to everyone but with all the traditional blocks(users have to find and install, different configurations might mess up your app etc)

Completely agree, I can't believe that Visual Basic was more advanced UI design-wise than current tools.

How about Dart + Polymer?

HTML+CSS+JS being the "assembly language of the web" is a terrible build target, I want a better foundation to build upon.

I think you're massively overstating how bad HTML+CSS+JS are, but how do you feel about WebAsm + WebGL as the target?

I'd like to keep the DOM around and not have the web turn into an incomprehensible mess of canvas pixels.

The DOM is perfectly appropriate for modeling documents, and we should indeed keep it around and make use of it where appropriate, but documents aren't all we want to do with the web platform.

A non-cloud off-site backup appliance.

This is what I want (I know there are alternatives, but this is what I want):

A device that I buy, and can plug at least one hard drive in to. I give it some sort of passphrase. I then place it in a friend's house and connect it to their internet connection.

I can then access it remotely from my house. I can easily backup my stuff to it. My backups are encrypted, both over the wire and on the drive.

When my house burns down, I can drive over there and get all my photos and records and stuff, instantly.

I don't pay a monthly fee. I buy the device and then it is mine.

I know I can build this myself. I don't have time. If I could buy this I would.

It would be nice if it was easy to use the other way around, e.g. buy one for your parents and keep it at your house and automate their backups somehow.

Hey there,

I know it does not totally answer your request, but you totally can achieve that with Crashplan; and for free. Simply buy a NAS or equivalent, place it at your parents and install Crashplan on it.

It is not as perfect as a totally untied system like you want, but it is free and offsite :).

I am a happy customer of Crashplan for a while but other than that I have no ties there :).


Yeah, I think it is close but not quite. It seems that you can back up to a friend's trusted computer, but it has to be turned on. Plus I don't necessarily want to trust them in that way.

If you expect your friend to accept having something plugged all the time, I assume you'd be ok with having a similar device at your place running their backup, right ?

In that case, you can buy a NAS for you, have them buy one for them, backup stuff on your own NAS and let the replication from your NAS to their NAS happen in the background.

Bonus: backups/restore/browsing are much faster, and you can more easily accommodate multiple people wanting to backup at the same place -- instead of one device at your place for each of your friend, it's only one at your place to which all your friend's NAS backup to.

I use Synology NAS, not with this setup, but I know they have the software to do exactly this.

Try Synology Hyperdrive. I'd assume that all the other NAS vendors do something similar.

My Synology NAS does exactly this. Besides this it backs up my Google Drive and Dropbox too.

You could probably do something like that with FreeNAS and ZFS send.

Also, Crashplan has a "backup to friends" feature.

Raspberry running an SSH daemon or rsync or attic or whatever?

There's been a few services like this. None have seems to have made it, at least I can't remember any names.

Product already exists in the form of consumer NAS devices. You don't have to build anything, just turn on remote access.

You can do this with Time Capsule I believe. Doesn't help for windows or linux of course.

I want exactly the same thing!

Instead of building, strengthen the independent pipe which is responsible for flowing the data/information across the Internet. I am talking about RSS. I wish more technical and policy people would consider supporting, or reviving, the RSS. RSS is practically not owned by anyone (like how email flows from one platform to another without ownership restrictions). The modern API world has proliferated silos and boundaries, which is ridiculous.

After weeks of Microsoft bouncing every mail our server sent to any of their mail properties (due to, I guess, some history on the IP before we owned it), I have begun to doubt the premise that no one owns email. The major providers can make your mail server useless to 20%, or more, of email recipients. That's a pretty big club to wield. (I understand why a mail provider would block, and I've instituted IP-based blocks on my own mail servers in the past. It's just that email is such a mess, and there's not a good way to solve it.)

But, I would love it if RSS made a huge comeback. Twitter, facebook, Instagram, etc., I'd love it if I didn't have to open any of them, but could still follow my friends posts. There's no technical reason for them to be walled gardens, only business reasons, which are at odds with my privacy and general happiness.

> There's no technical reason for them to be walled gardens, only business reasons, which are at odds with my privacy and general happiness.

Then pay them! Hosting fees are not gratis.

Where should I sign up for the "privacy-respecting, no-advertising" facebook plan? Twitter have one of those, too?

Nope, but google is trying. See google contributor and youtube red. Start there.

Google Contributor looks neat, and probably something I'll try out...but, I don't see any indication that privacy will be respected with their plan. Seems like I still need an ad blocker and a privacy-respecting web browser.

It's unfortunate that something like Contributor can't be done in a peer-to-peer fashion. It requires a huge player with a visible impact on their entire web to be able to "sell" a different sort of web experience. That's actually kinda scary; it reminds me how much power Google has. While I have vaguely positive feelings about Google (and use gmail, Android, etc.), it's not great that one entity owns such a big chunk of the web.

I already have a couple of subscriptions to services that provide music. I don't know if I'll switch to Red; might try it out at some point. I should read up on it; I do watch a lot of tech videos and such on YouTube, and if Red actually supports the people who make them, that'd be great.

Youtube Red and Google Play Music come in a package, which is actually an awesome value. Red does support the creators whose videos you watch.

Not OP, but unfortunately Google Contributor is not yet available in my country. Google really seems to be focused on the US and only on the US.

Yes. I am constantly telling people to switch to email after having contacted them on Facebook, WhatsApp and Quora. It would be nice if there is some enforced standard to exchange personally written information. Why not return to email? Perhaps upgrade email, for example by using utf-8? I have no idea, however, how to create a compatible upgrade path. This would be my wish: Someone should create a feasible upgrade path to email and create a standard which all mayor player have to follow.

I asked somewhere here, if there is a way to extract the email addresses out of Facebook (or Twitter), of friends etc. Apparently there is no solution (probably because these locked walled-gardens do not allow such extraction, even when you have first layer of trust established, i.e, I am only talking about extracting the email addresses of friends on Facebook).

So the gist is, that it is not really technically easy to have people move back to email's personal communication. I gave up on Facebook and I am not on any of these other social networks. I'd rather just communicate with people via email or personal text.

A great-looking, affordable, modular home kit. We did it 100 years ago with the Sears houses -- imagine what we could do nowadays with mass production and 3-d printing? Imagine if you could order a home kit and save half of what a traditional builder would charge you? How would that change opportunities for the middle class?

In a similar, Maslow's-Hierarchy-of-Needs vein, it would be great to have an app that crowdsources data about healthcare costs and other data points in your city. I'd like to know which hospital charges least for an M.R.I., which hospital has the highest rate of MRSA infections, which doctors are highest-rated by their patients, which insurance policy is the best in my location. Right now there's little-to-no transparency and, just like in Vegas, the House always wins.

In general, I would love to see tech take on disruption and increased affordability in the areas of true life needs -- affordable education, housing, medical care, healthy food -- and focus less on gaining tiny efficiencies in tools and workflows.

TL;DR -- I need an affordable home, not a refrigerator that sends text messages to my blender.

You likely wouldn't save half, I would guess ~20-30% provided you do the majority of the assembly. The era of Sears homes benefitted from lax or non-existent building code (especially for electrical/plumbing).

You may also have problems selling the house in the future being it was not built by a professional. In the near term you may be unable to get financing (if you do, it will likely be much higher than a standard mortgage rate) as there is no contract or guarantee the house will be completed; they are just paying for materials.

One of the biggest costs will be the property. You need a lot zoned for residential housing with available utilities (sewer, water, gas, electric) or alternatives (septic, well, propane, wind/solar).

The lot has to be surveyed, plans approved by the building commission and permits paid for.

The foundation plan will have to be drawn up by an engineer or architect because frost depths and local regulations vary. An owner could maybe pour a slab but likely couldn't do a basement on their own; either is best left to professionals.

Now the person has to assemble the house. You could get by doing much of the work alone but would likely need a helper for various stages throughout the project.

Most of the materials are cut but you still need some tools which would be another ~$2000+ expense; ladders, air compressor, nailers (framing, trim, roofing), drills.

When it comes to the roof, especially if a 2 story house you will have no choice other than hiring a crane to set the trusses. You'd likely have to hire a crew or at least have someone experienced working with you because this part is dangerous even for professionals. After the trusses are in place you could sheath/shingle the roof yourself.

Once the shell in complete you can move inside to finish the house. This is where your local laws will make the biggest difference in price. Some will allow homeowners to run plumbing and electrical themselves, others require all work be completed by a licensed professional.

After all the systems are installed (electrical, plumbing, gas, hvac) you can start finishing the house. This part is most conducive to the DIY process as it is all aesthetic. Drywall, flooring, trim, kitchens/baths.

TL;DR -- It would still cost a significant amount and require a huge amount of labor from both the owner/builder and professionals.

It's amazing how complicated and expensive we've made something that is relatively simple.

I've built most of a couple of stick-built houses and additions (fortunately in areas that have less insane regulatory codes), from the foundation up. It's really not that hard to do right.

It would just make the price of land go up proportionally

A way to use Mechanical Turk from outside the US. I've needed it with every client I ever worked with, and I need it now and know a dozen companies who do.

CrowdFlower has a de facto monopoly on the outside of the US supply and charges an enormous premium for it - enough to turn off most of them. I certainly don't want to pay several thousand a month for the right to submit jobs, although I was ok with the 25% premium in the old days.

If it's an alternative marketplace, it has to have excellent automation via API. I'm not going to use this for questionnaires, I'll be submitting thousands of jobs automatically.

If you have built this already, please email me.

Hey there—I'm CEO and cofounder of Scale API (YC S16, https://scaleapi.com), which is a developer-first API for human tasks. We allow customers outside the US, and are much more focused on quality than MTurk.

Definitely give us a shot, and feel free to email me if you have any questions! alex@scaleapi.com

Hi Alex, thanks for your email. Replying here for the benefit of other readers.

We can't use your product for the same reason we can't use most of the alternatives on the market: there's no way to push tasks with custom HTML/CSS templates, as is possible with MTurk. This is important at scale.

Do let me know if this ever changes.

Check out https://www.scaleapi.com/ not sure if they work outside of the US - but they are a MT alternative

Out of curiosity, what do you use it for? I'm intrigued by platforms like MT, but have never had a real use myself.

For my clients, the standard use cases for e-commerce.

For example, you might have 100,000 products plus another 5,000 new ones listed per month. Your human error rate might be 3%. When a product arrives, its buyer (the corporate person responsible for signing on the brand) gives it a category and it shows up on the site. So 3% of the new products are erroneously categorised by their buyer.

You give X human votes per product, say, 10. "Which of these categories applies to this product?"

If the votes agree with each other, you can just input that as the category. If there's a split, you can feed those products to a more open ended question like "what category is this product" and use that as a starting point for either renaming the categories, finding a new category, or tagging: if it's a 50/50 split, you can just tag the product with both categories.

This is all automated and used to cost me around $200 per 5,000 products.

Another example might be training a data set for a machine learning algorithm. You send a data set to be trained by Turk and use it as training/test sets. I'm keeping this one deliberately vague for now as AFAIK we are the only ones in our space doing this and I don't even want to mention the space due to a relatively smart competitor.

As a hypothetical example, you might be trying to predict the category from the description, gender, picture features, buyer and brand, and the above categorisation tasks can be used to train the algorithms you're testing out.

Interesting! Thanks


You can try Yandex.Toloka. It has API (unfortunately not compatible with MT)

Thank you. This is the best available solution and I didn't know about it.

We'll probably go with it for now, until we hit a scale of tasks that justifies setting up a US subsidiary and dealing with the IRS.

I've used Fancy Hands (https://www.fancyhands.com/) before and had good results with it, though it's probably not really price-competitive with MT; I paid about $15 for a quite large amount of scraping work that would've been annoying to write software to do.

Looks like a neat site, but it doesn't allow us to structure the task (use an HTML template). The flexibility is important.

I could not agree more. It's quite ridiculous that there's no non-US alternative that's affordable.

Do you use custom software to interact with MTurk, or is there already some good software for interacting with it already? I have an image dataset I want classified and I've been putting it off for a while...

Sites like CrowdFlower have point-and-click interfaces that allow you to just upload a CSV and drag and drop its columns as fields in your task template.

What I used to do was host the images on a server, publicly, then upload the URL of the images and use it to display the images within the task. It's OK for one-offs. You get a thousand free rows with the CrowdFlower trial [1] and that's probably the most painless way to go for a classification problem.

In our case though, everything will be written in-house. Automation is how we keep headcount low.

[1] https://www.crowdflower.com/plans/

I've used https://microworkers.com/ before (many years ago) and seems to do the job quite well.

How do they region lock it? Can't you just make a US business entity and pay through that, if the origin of payment is the issue?

I could, but it's time consuming and I'd have to deal with the IRS.

By not operating in either the US or the EU, I save myself a lot of administrative trouble and potential tax liabilities.

So you essentially just want a broker between yourself and MT? If all I have to do is provide you an account, and you pay up front with bitcoin, I'm happy to do this for you with 25% premium. All it involves for me is going out and buying prepaid debit cards of whatever amount and attaching them to the account. I fill them up with 75% of whatever dollar amount you send me.

Thanks for offering, but I doubt this methodology will scale. We really need a stable company with a product we can trust will both be around in 5 years and that can handle large volumes of tasks.

A gmail clone with privacy monetized via charging me $50 or so a year. I'm currently a fastmail subscriber, and it is nowhere near as good as gmail.

Fastmail specifically is deficient in several ways:

* gmail conversations. it is threads done correctly. Fastmail half-does this but the seams peek through all over the place. Eg you don't have labels, you have actual folders and those two aren't the same at all.

* fastmail search is still mediocre, and is clearly intended to be used via their graphical menu rather than typing folder/label restrictions or other modifiers in the search box.

* A gmail style iphone + android app that works offline

* better polish throughout the app (eg: if something is incorrectly assigned as spam, when you say not spam, message routing rules don't apply to it. If you create a filter, you have no option to apply to existing messages. I could go on and on.)

* spam detection that works way better

Fastmail may eventually be what I want however. They've definitely improved over the last 2 years. Eg they used to use 2fa as a monetization source (10c or so per text message!) and have recently made gmail style 2fa free. They've also turned their settings UI from appallingly bad (it looked and felt like a very junior developer's first js project) to pretty good. Similarly with their rules routing engine.

Have you looked into Proton Mail? I use the android app and have had no issues post-beta. I also haven't gotten any spam whatsoever.

Id love to move away from Gmail but I haven't found another client that has the features I cant live without. Most importantly the prioritized inbox.

I wish someone would build an interactive teaching AI, (perhaps in a mathematical context at first). For example, the user might start with a goal such as, "I'd like to understand singular value decomposition." The AI would interactively assess the user's level of background and begin instruction at the appropriate level, leading to the desired goal.

See Neal Stephenson, the diamond Age. Yes we need a tutor.

Furthermore many times this topic was discussed here on hackernews, with good insights and further links. Now how to find those?

I have the idea, we spent a lot of time struggling to progress on a particular topic and most of the time it just slips away. Can you provide a concrete example of what you would want to learn, what is your current skill level and what this service should do to provide value to you?

Ive always thought that the ship computer in Star Trek: TNG was a good example of how a collaborative AI should act.

arbital.com is close to that.

The AI starts getting mad.

You: I don't understand AI:. What is so difficult about quantum mechanics! I can solve this question in 0.00076 of a second!

Judgment day happens.

Maybe not so technical but as a curious person i feel i collect so many articles in instapaper but never take the time to reread them.

I do read though, mostly offline. Books, newspapers and articles. My focus is better offline.

Combining data mining and my offline focus, it would be great if you could make a tool that categorizes my instapaper articles and converts it to theme numbers (big data, health care, artificial intelligence, food, etc) that can be downloaded in pdf.

Would be so cool! Thanks for asking the question.

Or! Sent periodically as a custom magazine when the theme has received enough articles / reach a volume threshold to justify the print and sending.

I would like a vi for spreadsheets. It should run in the terminal, and be able to read/save csv/xls/xlsx/ods files. It should understand formulas. I should let me navigate with vi-style keys, and have a separate mode for editing a cell's contents (with escape to go back to navigation mode). It should perform well even for very large documents (many rows or many columns, but especially many rows). I propose that it be called `vc`. :-)

I gave up on sc when I realised there is no npv function that accept a cell range as cash flow input. I also wish it supports having different sheets per document.

I'd love that! Having a terminal UI for graphs too!

Emacs + orgmode if you want something today.

+ viper mode if you really want vi style modal editing

edited to add: or apparently spacemacs

A people-focussed email client.

Pretty much all email clients are message/thread based.

I'd rather have them people-based: A sidebar that shows a list of people who I have recently interacted with; and clicking on them shows all messages I've exchanged with them.

I don't really care about threads, because people just don't know how to use them. But even if they do, it often isn't clear when to start a new thread vs. continue an old thread.

Basically I want email to work more like an instant messenger app.

What OS? The only one I've heard of is Unibox (for macOS): https://www.uniboxapp.com/

Right, I completely forgot about this! I even bought Unibox on the Mac App Store when it came out! But while I loved the people centric concept, I just didn't like the design...

EDIT: I just downloaded it again, and here's why I don't like Unibox: it has a very low density design. It looks like a fancy website, rather than a productivity app. Lots of space is wasted on frames and borders. UI elements appear and disappear as you move the mouse. It's definitely a modern app, but I seem to prefer more traditional apps...

Simple, user-serviceable appliances.

Not a fridge with an UHD screen, not a washer with Bluetooth support, not a toaster that talks to the cloud.

Just functional appliances with a level of efficiency that existed 25-30 years ago and can be repaired, rather than thrown away because subcomponents are sealed black boxes with little regard to durability.

And after that, the really hard work: doing the same with printers.

I agree about simpler, durable, repairable appliances. Eg, http://www.jamesdysonaward.org/projects/lincrevable/

One important aspect is to ditch the electronic interface (buttons, LCD, etc), since it's often the weak point, both for durability and usability.

But IMHO you still need electronic control, for function/efficiency (timers, complex wash cycles, PID temperature control, dirt sensors, etc, etc), as well as usability.

However the UI should be replaced with a single on/off button, and bluetooth; the complex interface becomes a smartphone app or web page, which can be upgraded, hacked, and is in any case much more usable than LCDs, buttons and poorly designed constricted UI. And if there were a few cheap (super-mass-produced) general-purpose standard controller boards in use, rather than each manufacturer/model having a custom board, then repair/replacement of the electronics would be easy too.

> However the UI should be replaced with a single on/off button, and bluetooth; the complex interface becomes a smartphone app or web page,

As someone who has played the Mega Man Battle Network series [1], I love this idea.

In those games (highly recommended and well worth getting an emulator for), every appliance and machine basically has a universal interface which your "Navis" (think anthropomorphized avatars of Siri) can "jack into" and interact with.

In the real world, I guess something like that could be implemented as:

- Every appliance comes with an standard interface which exposes all its controls and configurable parameters.

- The first time you unbox a new appliance, you register it with your control device (computer/phone/watch.)

- After that you just use any app on your control device (like the HomeKit one on iOS) which supports the standard protocol, to enumerate and view each appliance's controls.

- There could be different levels of access depending on authentication and proximity. Say, a web page might only show you the basic status of all your appliances, but being on the same local network will offer extra controls, while physical contact between your phone/watch and an appliance via NFC will reveal its most sensitive settings.

[1]: https://en.wikipedia.org/wiki/Mega_Man_Battle_Network

For the most part, I think this is already the case today with a lot of products. Most people just have no idea a compressor works on a fridge or how to diagnose the problem. There's really nothing you could do to make a compressor less complicated to work with. Washers are driers don't exactly have embedded computers. They're straight forward to fix, too. Plenty of dishwashers available that are non-electronically controlled.

Most everything is fixable if you're not afraid of a screw driver and watching a Youtube on how to take it apart.

I agree with you here. I used to believe that all the appliances are not fixable due to all the comments people make like "they are not like they were 30 years ago".

In reality when I opened up some of my appliances for DIY repair I found them to be incredibly simple and easy to repair after some youtube videos. I am sure there are some examples where an appliance is not fixable but all the appliances I have owned in my life have been.

A smart bank account:

I want to set up rules like "take 19 percent of every incoming transaction and save it to virtual account 'taxes'. Use this, account to pay invoices by $financialAgency"

Banking hasn't produced any innovation since online banking, it seems

I guess you're not based in the UK, but monzo(previously Mondo) (https://monzo.com/ ) sounds like the start of what you're looking for. Their API (https://monzo.com/docs/) opens them up to all sorts of uses for your bank account.

I don't believe they have a full banking license yet do they?

No, but we are led to believe this wont be long. They recently got partial approval with full approval expected next year. Not bad for a startup that has only been going ~20 months in quite a slow industry.

Money is currently stored with a third party that do have a license, so there are some guarantees that money wont be lost.

Germany based, sadly. But I'll keep an eye on them, thanks

Jeez absolutely. The lack of innovation in customer services in banking is awful.

A bank account wherein I can ring-fence amounts and set a preferential order in which these ringed resources are drained. All with alerts on top. "Warning: you're spending more than recommended on fancy breakfast items, jeopardising your ability to pay your AWS monthly bill (which by the way has increased over the past 3 months)".

I feel like digital wallets would definitely solve this but there is never enough demand from consumers or merchants to give them sufficient traction. Maybe it's just a really slow shift since the benefits are marginal for most people.

Google Wallet recently added integration to bank accounts though so at least there are some progress in that aspect.

I'm watching Monzo (previously Mondo) with interest.

You should be able to do this with a HBCI client like Hibiscus/Jameica or similar.

I'll look into that, however this means additional software on my systems. I think it really should be integrated into the normal account and work without me interfering

I want Jarvis, literally the AI in the Iron Man movie. I want him to converse banter and operate my computers / servers with whatever I need, and do it FAST. Siri and Google are doing good things, but I want something I can host myself so its quick, responsive and feels like i'm talking to a real person.


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact